Statut : Manuscrit prêt à soumettre Contribution des auteurs :
L’auteur principal (Maeva Bavoux) : Design du système microfluidique, micro-fabrication, formation de sphéroïdes, traitement, analyse par tests de survie clonogénique et comet assays, analyses statistiques des résultats, rédaction de l’article scientifique, assistance à Yuji Kamio dans le design du support pour l’orthovoltage et pendant les expériences d’irradiation.
Le second auteur (Yuji Kamio) : Développement de la méthodologie pour l’irradiation avec l’orthovoltage, calibration des temps d’expositions pour obtenir les doses souhaitées, analyse de la dose reçue avec les films Gafchromiques, relecture du manuscrit final.
Le troisième auteur (Emmanuelle Vigneux-Folley) : Développement du plug-in pour l’analyse automatique de la taille des sphéroïdes dans le système microfluidique.
Les auteurs Julie Lafontaine, Ouafa Najyb, Elena Refet : Assistance dans les planifications expérimentales et relecture du manuscrit final.
Shririn Abbasinejad : Supervision de Yuji Kamio dans la réalisation de la méthodologie d’irradiation avec l’appareil d’orthovoltage.
31
Microfluidic system and kilovoltage X-ray dose modulation technique for the investigation of novel radiosensitizers and radioprotectors
Maeva Bavoux1,2,3, Yuji Kamio4,7, Emmanuelle Vigneux-Folley5, Julie Lafontaine2,3, Ouafa
Najyb2,3, Elena Refet2,3,6, Shirin Abbasinejad Enger7, Thomas Gervais*2,3,5,6 and Philip Wong*2,3,4
1Department of Pharmacology and Physiology, Université de Montréal, Montréal, Qc, Canada 2Institut du Cancer de Montréal, Montréal, Qc, Canada
3Centre de Recherche du Centre Hospitalier de l’Université de Montréal (CRCHUM), Montréal, Qc, Canada
4Department of Radiation Oncology, Centre Hospitalier de l’Université de Montréal (CRCHUM), Montréal, Qc, Canada
5Department of Engineering Physics, École Polytechnique de Montréal, Montréal, Qc, Canada 6Institute of Biomedical Engineering, École Polytechnique de Montréal, Montréal, Qc, Canada 7Department of Biomedical Engineering, McGill University, Montréal, Qc, Canada
Corresponding author: Philip Wong.
Centre Hospitalier de l’Université de Montréal. 900 rue Saint Denis, Montréal, Qc, Canada. Phone: +1 514-890-8254
Email: philip.wong.chum@ssss.gouv.qc.ca
Key words: Radiotherapy, Microfluidic, Drug, Radiosensitizers, Radioprotectors, Repurposing
32
4.1 - Abstract
Radioresistance, tumor microenvironment and high toxicity of radiation limit the efficacy of radiotherapy (RT) in treating cancers. This challenge can be tackled by the discovery of new radiosensitizing and radioprotecting agents aiming at increasing the therapeutic index of radiotherapy. Here, we developed a novel microfluidic system capable of growing 336 homogeneous spheroids of soft tissue sarcoma (STS) while preventing bystander effects, and screened approved drugs in combination with RT. An orthovoltage-based dose-modulation technique was used to expose the systems to several radiation doses increasing the experimental throughput. Radiation dose dependent DNA double strand breaks in STS spheroids were observed using comet assays. Analysis of proliferative death using clonogenic assay was performed, and synergy quantified. Using this framework, we demonstrated that a PARP inhibitor, Talazoparib, has radiosensitizing properties on STS when combined with 2 Gy of RT.
33
4.2 - Introduction
At least 60% of cancer patients will receive radiotherapy (RT) as part of their treatment1. RT damages DNA directly by ionising DNA, or indirectly through the ionisation of intracellular molecules, production of reactive oxygen species and their subsequent interaction with DNA2. These DNA damages include single strand breaks (SSB) and double strand breaks (DSB), which induce various cell death pathways such as apoptosis, necrosis and irreversible cell cycle arrest (senescence).
Soft tissue sarcomas are rare cancers affecting less than 1% of the population. Because of the rarity and the heterogeneity of these cancers, treatment is challenging3. The main modality of treatment consists of surgery with adjunctive radiotherapy4. The efficacy of RT in treating cancers is limited by inherent tumor radioresistance, tumor microenvironment, and the radiosensitivity of normal tissues surrounding the tumors5. Over the last decades, engineering and imaging advancements vastly improved the accuracy of RT, thereby allowing the delivery of higher RT intensity and improving RT's therapeutic potential. However, few radiosensitizers or radioprotectors, aimed at increasing the therapeutic index of RT through biological manipulations, have been identified6.
In recent years, pharmaceutical industry has focused on drug repurposing to accelerate the discovery of new radiosensitizers and radioprotectors7. Drug repurposing involves evaluating known molecules (approved or abandoned) for new medical indications other than the one they were originally designed or approved for. As drug design, optimisation and toxicity studies of these molecules are often completed, drug repurposing cuts the time in drug discovery as well as costs by valuing research done to develop molecules that failed clinical trials8,9.
Indeed, 95% of drugs demonstrating anticancer activities in preclinical studies, ultimately fail in clinical trials due to insufficient efficacy or high toxicity10. Conventional drug screenings are usually done on cell monolayers, which are now recognized as poor predictors of drug efficacy11– 13. Due to the limitations of 2D cancer models, academic researchers and pharmaceutical companies are turning to 3D cellular models14. Several systems have been developed for drug screening on spheroids, especially microfluidic systems15–20. Miniaturisation of spheroid
34
formation using microfluidics offer several advantages such as reagents economy and process simplification. However, they have not been optimized for the combinatorial screening of drugs with RT. For example, irradiated cells can communicate with unirradiated cells to induce bystander effects in cellular proliferation and survival, and thus spheroids exposed to different radiation conditions must be isolated21,22. Another experimental challenge, which limits throughput, involves the inability to administer different RT doses to various samples within the same plate or system treated with conventional irradiators.
We have developed a microfluidic system to test combinations of radiotherapy and chemotherapy on spheroids23. A methodology using the 6 MV beam of a 21EX Varian linear accelerator was developed enabling the irradiation of a system with three different doses homogeneously in every other chamber, utilizing the chambers in between to ramp up the dose. In this study, we optimized the system and irradiation technique by first isolating the culture chambers to prevent bystander effects, and secondly by increasing the number of RT conditions that could be tested per system. A framework was developed to analyze the radiosensitizing and radioprotecting potential of drugs by clonogenic assays using a new orthovoltage-based dose- modulation technique to irradiate our microfluidic system.
35
Figure 1: Proposed approach to screen drugs with radiotherapy on spheroids. 1) Spheroids are
formed inside the microfluidic system. 2) Spheroids are exposed to drug-RT combinations. 3) Proliferative death following the combinatorial treatment is analyzed with clonogenic assays of dissociated spheroids. 4) Synergy quantification analysis of clonogenic survival values is performed using dose modifying ratios (DMR) and combination indices (CI).
36
4.3 - Material and Methods
Design of the microfluidic systems
Microfluidic systems were designed using CATIA (Dassault Systemes, France) and made of polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning, Midland, USA). The design was made following design rules to ensure proper nutrient supply24. The medium change must be done up to 29 hours. A magnetic valve system was integrated in the systems to isolate the culture chambers. The design of the magnetic valve system was done following a previously published study19. Briefly, channels are closed using a metallic rod which, when attracted to a magnet, deforms a thin PDMS membrane, closing the channels. The system is composed of three layers of elastomer. The bottom layer contains the channels and micro-wells for spheroids formation. The upper layer contains oval holes where metallic rod can be inserted. The middle layer is a 100 μm thick flexible PDMS membrane.
Microfluidic systems fabrication
Molds for the upper and bottom layer of the systems were made of poly-methyl methacrylate (PMMA) using a Computer Numerical Control (CNC) micromilling machine (Roland MDX-40A, Irvine, California, USA). The two layers were then formed by pouring a mix of PDMS with a curing agent at a 10:1 mixing ratio into the molds and by curing them at 80°C for 2h. The 100 μm PDMS membrane was fabricated using a previously published protocol19. Briefly, the membrane is formed by pouring PDMS mixed at a 20:1 ratio over a silanized glass wafer. Then, the wafer is spin-coated using a Cee 200CBS (Brewer Science, Rolla, Missouri, USA) at 300 rpm for 60 s and cured at room temperature for 72h. Assembly of the device was done by first connecting the bottom layer to the membrane using oxygen plasma treatment, and then to the upper layer. The magnet-holder was 3D-printed. Twelve magnets (B444-N52, K&J Magnetics, Pipersville, Pennsylvania, USA) fit in the holder to actuate the magnetic valve array.
37 Spheroid formation and observation
Two primary human STS cell lines were used for the experiments (STS117 and STS93), which were previously characterized and described25. Cells were cultured in DMEM:F12 (Thermo Fisher, Ontario, Canada) with 10% bovine serum (Sigma-Aldrich, Ontario, Canada) and 1% pen-strep (Thermo Fisher, Ontario, Canada). Bubble-free filling was achieved by using 100% ethanol, then three washes of 70% ethanol were performed to sterilize the channels. To prevent cell adhesion to the PDMS, the channels were passivated using Pluronic (10 mg/mL, Pluronic® F-108, 542342, Sigma-Aldrich, USA) and incubated overnight in a humidity chamber at 37°C. Afterwards, channels were rinsed three times, with 5 minutes incubation between each rinse using 70% ethanol, HBSS (Hank’s Balanced Salt Solution, Thermo Fisher, Ontario, Canada) and finally filled with DMEM:F12. Spheroids were formed by introduction a solution of 2x106 cells/mL (STS117 or STS93 cells) in the channels. Ninety μL of cell suspension was introduced in the plastic inlet and 90 μL were quickly removed in the outlet. This process was repeated 3 times in each direction to ensure uniform cell distribution in the channels. The microfluidic systems were incubated at 37°C and 5% O2 in an autoclaved pipette tip box with a humidified paper tissue. Medium was changed every 24h. Bright field images were taken once the spheroids formed, two days post seeding.
An ImageJ (Wayne Rasband, National Institutes of Health, Bethesda, Maryland, USA) plug-in was developed to automatically analyze spheroids diameter. Images were segmented by the following consecutive steps: application of a median blur, threshold operation and contour finding. Areas are converted from pixels to microns and diameters are determined assuming the spheroids are circular. An Excel spreadsheet containing the diameters of all spheroids was exported. Spheroids diameters in the different chambers are presented as means ± SD for three experiments with 84 spheroids analyzed per replicate.
Irradiation using orthovoltage radiation
A dose-modulation technique using an Xstrahl 150 unit (Xstrahl Inc., Georgia, USA) was developed. The beam energy was 140 kV (HVL = 8.42 mm Al) with additional filtration (1.15 mm Al + 0.2 mm Cu). The field size was collimated by a 15 cm cone (SSD = 25 cm). The beam was calibrated following the ‘’in air’’ method of AAPM TG-61 protocol26 using an A12 farmer chamber
38
(Standard Imaging, Wisconsin, USA) with air-kerma calibration factor [cGy/nC] traceable to a national metrology institute standard (NRC, Ottawa, Canada). A support (80% PLA) was printed with an Ultimaker S5 3D printer (Ultimaker, Geldermalsen, Netherlands) and Fusion 360 CAD software (Autodesk Inc., California, USA) to reproducibly place the microfluidic system within the field with 4 notches that secures the system to the cone. In addition, the support includes a slit allowing moving a 3 mm lead collimator plate used to modulate dose-levels using a step-and- shoot technique. Radiation dose increments were achieved by moving the lead plate between three irradiation fields. The dose-rate at the microfluidic system surface was calculated to be 1.71 Gy/min (with a tube current 10 mA) after correcting for an air gap of 1.7 cm. Calculated exposures times were subsequently optimized to yield the desired dose-levels (8 Gy, 4 Gy, 2 Gy and 0.5 Gy) based on EBT3 Gafchromic TM film results, namely 2.8 min, 1.13 min and 0.9 min for the first, second and third irradiation fields performed as a single fraction. EBT3 Gafchromic TM films were analysed using a triple-channel dosimetry method27. A detailed explanation of the irradiation technique and its validation has been published28.
Patra et al.23 microfluidic systems were irradiated using a cell irradiator (Gammacell 220, Atomic Energy of Canada Ltd, Ontario, Canada). A single radiation dose per system was delivered. Drug treatment
Talazoparib (BMN-673) was purchased from MedChem Express (Monmouth Junction, New Jersey, USA). Pazopanib and AZD7762 were purchased from Selleckchem (Houston, Texas, USA). Talazoparib, Pazopanib and AZD7762 were diluted in culture medium at various concentrations. The treatment started two days post-seeding once spheroids were formed. Irradiation was performed 24 hours after treatment induction. Spheroid collection for comet assays was done 30 minutes post-irradiation. Spheroids were collected for clonogenic assays 24 hours post- irradiation.
Clonogenic assay
The microfluidic systems were peeled, and spheroids were collected using PBS (Thermo Fisher, Ontario, Canada). Dissociation into single cells was performed by incubating the spheroids in Trypsin-ETDA 0.25% (Wisent, Quebec, Canada) for 7 minutes at 37°C. For each condition, cells
39
were counted and seeded in triplicates in 6-well plates: 500, 700, 1500 and 6000 cells for the non- irradiated, irradiated at 2 Gy, 4 Gy and 8 Gy conditions respectively. The plates were left 10 days in the incubator to allow for colony formation. Colonies were fixed and stained using 70% methanol solution containing 0.5% Crystal Violet (Sigma-Aldrich, California, USA). Percentage of survival fraction values were obtained by normalizing the number of colonies counted for each treated condition with the number of colonies in the control. Fractional survival values were obtained by normalizing survival fraction values to non-irradiated controls. Survival curves were fitted with the linear–quadratic model using GraphPad Prism 8.2.0 (Graph-Pad Software Inc, San Diego, California, USA). Combination indices from the Chou-Talalay model29 were computed with CompuSyn (T. C. Chou and N. Martin, Memorial Sloan-Kettering Cancer Center, New York, USA) for a non-constant ratio. Data are means ± SEM of three biological replicates.
Comet assay analysis
The neutral comet assay (Trevigen Inc, Gaithersburg, Maryland, USA) was performed to detect double strand breaks following irradiation. The spheroids were irradiated two days post-seeding. The comet assay was performed 30 minutes after irradiation. Spheroids were collected from the microfluidic systems using PBS and dissociated using Trypsin-EDTA 0.25% for 7 min. Cells were diluted in DMEMF12 to obtain a cellular suspension of 2-3x105 cells/mL. 90 μL of this solution was added to 300 μL of softened LMA agarose maintained at 37°C. An aliquot of 50 μL was spread onto each comet slide, which were kept at 4°C for 10 minutes to allow the agarose to solidify. Afterwards, slides were first transferred to a cold lysis buffer for one hour, then to a cold 1X Neutral Electrophoresis Buffer for 30 min. The slides were subjected to electrophoresis in a cold 1X Electrophoresis Buffer for 20 minutes at 21V, immersed in DNA precipitation solution for 30 min, in 70% ethanol for 30 min and finally air dried at 37°C for 30 min. Slides were mounted with ProlongTM Gold Antifade Mountant (Invitrogen, Grand Island, New York, USA) and DAPI (Sigma-
Aldrich, California, USA). Images were taken using an inverted microscope (Zeiss AxioObserver Z1, Carl Zeiss, Jena, Germany) and analyzed using CometScore 2.0 (TriTek Corp., Sumerduck, Virginia, USA). Tail moments are defined as: 𝑇𝑎𝑖𝑙 𝑚𝑜𝑚𝑒𝑛𝑡 =Tail lenght x % DNA in the tail
40
Data are means ± SEM for three biological replicates with more than 60 cells analyzed per replicate.
4.4 - Results
Validation of the microfluidic system and irradiation method
The microfluidic system was designed to test up to 16 conditions of drug-radiotherapy combinations (Fig. 2.A). To ensure that culture chambers are isolated, a magnetic valve system was integrated based on our previous work19. When attracted by a magnet, a magnetic rod deforms the PDMS membrane and thus close the channel (Fig. 2.B) between chambers. A graphic representation of the system’s operation is illustrated in Fig. 2.C.
Two days post-seeding, spheroids size-homogeneity inside the device was studied. Spheroid diameters were automatically segmented from bright field images using our custom ImageJ plug- in (Fig. 2.D). The diameter was computed from the area assuming spheroids are spherical. The mean diameter of STS117 and STS93 spheroids in each chamber was determined from three independent experiments (Fig. 2.E). No statistical differences were found in spheroid size between chambers for both cell lines (one-way ANOVA). The mean diameter of STS117 and STS93 spheroids was 324 μm ± 15 and 257 μm ± 19 respectively. The spheroids formed in the different culture chambers were homogeneous in size (SD < 15%).
Fig. 3.A illustrates the experimental setup of the irradiation process using an orthovoltage unit. The microfluidic system was positioned within the field using a 3D-printed support with 4 notches that secures the system to a 15 cm cone. A 3 mm lead plate was used to modulate the dose-levels. A radiochromic film was placed at the spheroids level (Fig. 3.B) to generate a dose map of the radiation dose received by the cells. Fig. 3.C shows the averaged multichannel dose map of the microfluidic system irradiated with four dose-levels: 0.5 Gy, 2 Gy, 4 Gy and 8 Gy. These radiation dose increments were achieved by moving the lead plate between three irradiation fields delivered as a single fraction. The total irradiation process takes less than 10 minutes. Unirradiated sections received on average 0.5 Gy. Fig. 3.D shows that in a regular radiotherapy dosimetry plan, normal tissues surrounding the tumor receive at least 0.5 Gy.
41
Figure 2: Microfluidic system design. A) 3D design of the microfluidic system and picture of one
culture chamber with STS93 (soft tissue sarcoma) spheroids. B) Cross-section of the microfluidic system and illustration of the magnetic valve system operation (adapted from R-Brunet et al19). Channels are closed using a metallic rod to deform a thin PDMS membrane. C) Example of the system operation. The two bottom channels were filled with ethanol and one was closed using a metallic rod. A red dye was subsequently added. Scale bar = 1 cm. D) Automated segmentation of spheroid diameters in a culture chamber using an ImageJ plugin (described in Materials and Methods). E) Mean diameter of spheroids in the four chambers of the microfluidic system (21 spheroids/chamber). Values are presented as mean ± SD of three independent experiments (n = 84 spheroids per group).
42
Figure 3: Dose-modulation technique using an orthovoltage beam to irradiate the microfluidic
systems. A) Schematic of the irradiation experimental setting. A 3D-printed support includes a slit allowing to move a 3 mm lead collimator plate. The lead plate is used to modulate dose-levels. B) Radiochromic film positioned between the two outer PDMS layers, at the spheroid-level. C) Dose map generated from scanned radiochromic film optical density values. D) Example of radiation doses distribution used in clinic to treat cancer patients. The color map displays radiation doses superior to 0.5 Gy when 8 Gy is aimed at the target (Green contour)
43 Radiation-induced DNA double-strand breaks analysis
Clonogenic cell survival assay is the gold standard to determine the radiosensitivity of cancer cells30. We examined the clonogenic cell survival of STS117 spheroids that were dissociated 24h post-irradiation (Fig. 4.A). Additionally, we investigated the molecular effect of RT on nuclear DNA using the neutral comet assay, which mainly quantifies DNA double strand breaks (DSB) that leads to RT-related cell deaths. Spheroids were irradiated on-chip with the four radiation doses, and the comet assays were performed 1h post-irradiation. The tail moment, defined as the product of the fraction of DNA in the tail and the tail length, is proportional to the number of double strand breaks31. In the fluorescence images of DAPI-stained comets (Fig. 4.B) we can see that the length of the tail increases with the radiation dose. Tail moments of STS117 and STS93 cells, irradiated as spheroids and then immediately dissociated, are shown in Fig. 4.C. Statistical differences were found between the tail moments of the control condition (0.5 Gy) and of spheroids exposed to high radiation doses (4 Gy and 8 Gy) for both cell lines. No statistical differences were found between the tail moments of the two cell lines (p>0.05, unpaired t-test). To assess the potential of the neutral comet assay to compare the radiosensitivity of a cell line, a correlation between STS117 clonogenic cell survival and tail moments was investigated (Fig. 4.D), yielding a high correlation (R2 = 0.98)32.
44
Figure 4: Radiosensitivity of sarcoma cell lines. A) Clonogenic assay of dissociated STS117
spheroids irradiated with four different doses. B) Quantification of DNA double-strand breaks of irradiated soft tissue sarcoma spheroids using the neutral comet assay. Representative fluorescence images of DAPI-stained comets. C) Average tail moment calculated with CometScore 2.0. The analysis was performed with at least 100 comets per condition. Values are presented as mean ± SEM of three independent experiments. A one-way ANOVA was used to determine the statistical significance between the different conditions (*p<0.05). D) Relationship between tail moments and clonogenic survival values of STS117 spheroids irradiated with different radiation doses. The data were fitted using a linear trend and the R2 value was computed.
45
Investigation of the potential of Talazoparib as a radiosensitizer for STS